Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. 2006 Nov 7; 103(45): 16840–16845.
Published online 2006 Oct 30. doi:  10.1073/pnas.0607904103
PMCID: PMC1636541

Template disruptions and failure of double Holliday junction dissolution during double-strand break repair in Drosophila BLM mutants


Previous biochemical studies of the BLM gene product have shown its ability in conjunction with topoisomerase IIIα to resolve double Holliday structures through a process called “dissolution.” This process could prevent crossing over during repair of double-strand breaks. We report an analysis of the Drosophila BLM gene, DmBlm, in the repair of double-strand breaks in the premeiotic germ line of Drosophila males. With a repair reporter construct, Rr3, and other genetic tools, we show that DmBlm mutants are defective for homologous repair but show a compensating increase in single-strand annealing. Increases of 40- to 50-fold in crossing over and flanking deletions also were seen. Perhaps most significantly, the template used for homologous repair in DmBlm mutants is itself subject to deletions and complex rearrangements. These template disruptions are indicative of failure to resolve double Holliday junctions. These findings, along with the demonstration that a weak allele of topoisomerase IIIα has some of the same defects as DmBlm, support the dissolution model. Finally, an analysis of DmBlm mutants in conjunction with mus81 or spnA (Rad51) reveals a second function of BLM distinct from the repair of induced double-strand breaks and possibly related to maintenance of replication forks.

Keywords: Bloom syndrome, DNA repair, mus81, synthesis dependent strand annealing, Repair reporter 3

Cells possess many ways of repairing double-stranded DNA breaks, and with good reason. Genomic integrity must be restored after the occurrence of a double-strand break (DSB) to avoid cell cycle arrest or apoptosis. Common sources of DSBs include the collapse of replication forks, transposable element excision as well as exogenous sources such as ionizing radiation and chemical mutagens. The mechanisms available for repair of DSBs are often classified according to whether homologous sequences are exploited (13). Homologous repair (HR) can make use of an identical but unbroken copy present on the sister chromatid, an outcome we shall refer to as HR-s, or it can use the homolog. The latter, which we term HR-h, can cause homozygosity at sites near the break. Homologous repair can also use direct repeats flanking the break, as in the single-strand annealing (SSA) mechanism (4). Nonhomologous end-joining (NHEJ), in contrast, requires no extensive homology but is more error-prone (5).

The mechanisms for DSB repair are evolutionarily ancient and highly conserved. Many of the genes have acquired a multiplicity of functions and appear to play multiple roles. A good example is BLM, the gene responsible for Bloom syndrome in humans (6), whose symptoms include greatly increased sister chromosome exchange and susceptibility to a wide range of cancer types (7). BLM is a member of the RecQ helicase family (8). Its biochemical activities include ATP-dependent Holliday junction migration, 3′-5′ strand displacement, and ATP-independent single-strand annealing (9, 10). Most significantly, BLM protein in conjunction with topoisomerase IIIα (Top3α) can resolve double Holliday junctions (dHJs) via dissolution or convergent migration (1114). This activity could allow BLM to suppress crossing over by resolving HR intermediates without strand exchange. Other roles proposed for BLM cover a wide range. The processes and conditions that have been implicated include spindle assembly checkpoint activation (15), telomere stability (16), recovery from replication fork collapse (1719), NHEJ in Drosophila (20), HR via the synthesis-dependent strand annealing (SDSA) pathway (21) which does not involve double Holliday junctions, regulation of chromosome synapsis (22), DNA damage signal transmission (19) and disrupting DNA displacement loops (“D loops”) to preclude formation of double-Holliday structures (23).

This plethora of biochemical activities and possible biological roles makes the analytical study of genes like BLM a particular challenge, especially in the context of multicellular organisms. One useful approach is to concentrate on a particular tissue type, such as the premeiotic germ line in Drosophila males, where a broad spectrum of relevant parameters can be measured. The lack of meiotic crossing over in Drosophila males and the potential for scoring large numbers of offspring from single individuals allows for precise quantitative measurements with impressive statistical power. Moreover, tools such as Repair reporter 3 (Rr3) allow simultaneous measurement of the relative usage of multiple DSB repair pathways (24, 25). Studies with Rr3 have shown that changes in the usage of one pathway tend to trigger compensatory changes in others, thus emphasizing the value of simultaneous measurements of multiple DSB repair outcomes (24).

The Drosophila BLM gene is mus309, also known as DmBlm (26, 27). Its known phenotypes are sensitivity to DSB-causing agents, female sterility, and increased frequency of deletion formation (26, 28, 29). It also has been found to cause a sharp decrease in a class of events that combine HR synthesis from the sister chromatid and collapse of a tandem duplication (21). Here, we use Rr3 and other tools to analyze the function of DmBlm in the Drosophila male germ line by measuring an array of parameters related to DSB repair. Our results are consistent with a model in which the primary function of BLM in repairing induced DSBs is the dissolution of double Holliday junction intermediates. The genetic analysis also separates this role from an essential function that can be performed by either BLM or Mus81.


Measurements of DSB Repair Pathway Usage with Rr3.

DSBs were induced by a rare-cutting endonuclease, I-SceI, in the premeiotic male germ line of Drosophila. The site of the cut was the repair reporter construct, Rr3 (24, 25). The scheme in Fig. 1A was used to measure the relative usage of SSA, NHEJ, and homologous repair from homolog (HR-h) to repair the breaks, along with the frequencies of recombination and flanking deletions. The results (Fig. 1 B and C) showed a sharp decrease in HR-h in DmBlm mutants relative to heterozygotes or WT. The DmBlm mutants used were homozygous-nulls but may have some maternal effect DmBlm product early in development (see Discussion). There was also a simultaneous increase in SSA, which roughly compensated for the lack of HR-h. There was no detectable change in NHEJ. Crossing over increased in the DmBlm mutants >50-fold relative to the heterozygotes or the wild-type controls. The DmBlm mutants also showed an increase in flanking deletions (Fig. 1C), similar to the finding of McVey et al. (29).

Fig. 1.
Tests of DSB repair with Rr3. (A) DSB repair pathways are measured in the germ line of individual males. In this example, the test males are null for DmBlm. All test males are heterozygous for the reporter construct, Rr3 (24). This construct includes ...

To assess the statistical significance of these differences, we use the distributions of measurements obtained from individual males. The variability between males comes from sampling variation and from premeiotic events, which can yield multiple sperm carrying the product of a single DSB repair. However, the estimate obtained from scoring the offspring of each individual male is independent of every other such estimate, thus providing considerable statistical power for detecting differences. Some of the distributions are shown in Fig. 6, which is published as supporting information on the PNAS web site. By applying the permutation test described in Methods, we see that the effect of DmBlm on HR-h, SSA, crossing over and flanking deletions were all significant at P < 10−6.

One interpretation of the results in Fig. 1B is that the decrease in HR-h is a direct consequence of the DmBlm mutation, whereas the offsetting increase in SSA is a secondary effect owing to competition between DSB repair pathways (24). Alternatively, the primary effect could be increasing SSA, with the drop in HR-h being a compensating consequence. To distinguish between these possibilities, we performed an experiment similar to Fig. 1A except that the template construct, Rr3EJ1, was not present. Therefore, there was no suitable template for HR-h. The results (Fig. 7, which is published as supporting information on the PNAS web site) showed that SSA was no longer affected by DmBlm, with frequencies of 79.9 ± 1.7% and 79.0 ± 1.5% in the mutants and heterozygous controls, respectively. The difference was not significant (P = 0.36). There was also no significant difference between the frequencies of NHEJ (22.2 ± 1.9% and 19.0 ± 1.4%, see Fig. 7). We conclude that the drop in HR-h is a direct consequence of loss of DmBlm function and that the compensating increase in SSA is secondary.

Because dHJ dissolution requires Top3α as well as BLM (11) it was of interest to analyze DSB repair with a Top3α mutant. Unlike DmBlm, Top3α null mutations are lethal in Drosophila (30), making them unsuitable for use with the Rr3 tests. However, a P element insertion 7 bases downstream of the first exon/intron junction, Top3αEP2272 (30, 31), is viable. We postulated that this allele might be hypomorphic, even though no effect on the Top3α level was detected by Plank et al. (30) in larvae or adults. The results of Rr3 tests in Fig. 1 B and C show that homozygotes for Top3αEP2272 displayed all of the same effects as DmBlm-null flies, although to a lesser extent. That is, they had a decrease in HR-h and increases in SSA, recombination, and flanking deletions relative to WT (P < 0.04). We attribute the relatively weak effect compared with DmBlm to using a hypomorph allele of Top3α rather than a null mutation.

Interaction Between DmBlm and mus81.

The structure-specific endonuclease, mus81, is thought to be involved in resolving Holliday junctions (3234). Double mutants of BLM and mus81 are synthetic lethals in Saccharomyces cerevisiae (35). To test this combination in Drosophila, we first generated a deletion of mus81 by making use of our observation (Fig. 1C) that flanking deletions during DSB repair occur much more frequently in DmBlm mutants. The resulting deletion allele, mus815.1, and its derivation are in Fig. 8, which is published as supporting information on the PNAS web site. Homozygotes of mus815.1were viable, fertile, and showed little, if any, hypersensitivity to the chemical mutagen methyl methanesolfonate (see Supporting Methods, which is published as supporting information on the PNAS web site).

The double mutant of DmBlm and mus81 was lethal (Fig. 2A), even though both mutations are viable individually. Thus, the synthetic lethality seen in yeast also occurs in Drosophila. Interestingly, most of the double mutants survived until the late larval stages, but died during pupation (Fig. 2A). This pattern of lethality is typical for genotypes that suffer severe unrepaired DNA damage, because most larval growth occurs by cell enlargement with polytenization rather than cell division (36). Therefore, we attribute the lethality of the double mutants during metamorphosis to loss of the imaginal disk and histoblast cells, which are actively cycling and susceptible to DNA damage.

Fig. 2.
Tests of mus81 phenotypes. (A) Synthetic lethality of mus81 and mus309. Survival frequency is measured from the numbers of Tubby and Tubby+ offspring from a cross in which both parents were homozygous (or hemizygous) for a null allele of mus81 on the ...

We used a cross similar to that in Fig. 2A to show that the triple mutant of DmBlm-mus81-spnA/Rad51 is also lethal. Therefore, the synthetic lethality of BLM and mus81 mutants is not rescued by a Rad51 mutation, even though the corresponding triple mutant is viable in yeast (37).

Fig. 2B and Fig. 9, which is published as supporting information on the PNAS web site, show the results of Rr3 tests with the homozygous-null mus81 males. There were no statistically significant differences between the mutants versus the WT controls for any of the parameters tested, despite large sample sizes. We conclude that mus81 by itself has no detectable effect on the repair of induced DSBs.

Effects of DmBlm on Recombination and Long-Gap HR-h.

We devised a second system to study how BLM affects homologous repair and recombination (Fig. 3A and B). This system, like Rr3, uses I-SceI to create a DSB. However, repair of this break via HR-h requires copying in of ≈5,000 additional base pairs from the homolog, as opposed to only 12 bp when Rr3 is used as in Fig. 1A. In other words, HR-h becomes equivalent to filling in a 5,000-bp gap. The system also allows us to monitor recombination frequencies on the left arm of chromosome 2, which does not contain the I-SceI cut site, as well as on the right arm, which does (Fig. 3B).

Fig. 3.
DSB formation opposite a large insertion. (A) Double-stranded 54-mer used to replace P{w*} insertion at cytological position 50C. The I-SceI recognition sequence flanked by 17-bp P element termini was inserted into chromosome 2 at the identical position ...

For the left arm, we found that DmBlm mutants had ≈1% recombination regardless of the presence or absence of I-SceI endonuclease or the I-SceI cut site (Fig. 3C). No such recombination was observed in the heterozygous controls. The difference between mutants and controls was significant in each of the five pairs examined (P < 10−6).

For the right arm, which includes the cut site, there was still significantly more recombination in the DmBlm mutants compared with heterozygous controls even in the absence of the endonuclease or the I-SceI cut site (P < 10−6 for all six comparisons). However, there was an additional increase in recombination in situations where an induced break was made (Fig. 3C). For example, when endonuclease was supplied zygotically, the recombination rate for DmBlm mutants was ≈10-fold greater when the I-SceI cut site was present versus when it was absent (P < 10−6).

We used the heterozygous molecular markers labeled L and R in Fig. 3B to examine in more detail a sample of 210 independent recombinants on the right arm. All of these recombinants came from the DmBlm mutants with zygotic endonuclease. We found that 80 of them were either LC-RA or LA-RC, implying crossing over occurred within a 3.7-kb span that includes the cut site. The others can represent cases in which the conversion tract included one or both markers, or crossovers elsewhere on the right arm of the chromosome and independent of the I-SceI cut site. Interestingly, only 47 of the 80 had both P{CaSpeR} ends intact. The other 33 had a deletion and/or part of the I-SceI cut site. We interpret these 33 cases as analogous to the event we describe below as “template disruption” and infer that a similar phenomenon can occur in recombinant as well as nonrecombinant chromosomes.

The HR-h frequencies in Fig. 4A confirm our previous result that HR-h is greatly reduced in DmBlm mutants. The difference between the mutants and heterozygous controls was significant at P < 10−6, when endonuclease was supplied zygotically or through both maternal effect and zygotic genotype. The reduction in HR-h relative to controls was comparable to that observed with Rr3 (Fig. 1B) despite the much-increased gap length.

Fig. 4.
HR-h and template disruptions. (A) HR-h and template disruptions were measured by the scheme in Fig. 3 A and B. HR-h was computed as the proportion of offspring expressing w* among those receiving the cn+ bw+ chromosome. The template disruption frequency ...

Previous data showed that the I-SceI endonuclease used here can produce DSBs in the germ line via maternal effect (24). However, Fig. 4A shows no significant reduction of HR-h in DmBlm mutants when endonuclease was supplied only via maternal effect (P = 0.28). Our interpretation (see Discussion) is that some DmBlm product is also available early in development to act on these early DSBs.

Template Disruptions.

The system in Fig. 3B allowed us to detect a different and unexpected kind of event which we call template disruption. These events are cases in which the homolog lacking the cut site, i.e., the one marked with cn and bw, was itself altered during DSB repair. Such events are indicated by loss of white* expression. As Fig. 4A shows, we found significant production of template disruptions in DmBlm mutants, with these events composing nearly 1% of the progeny when endonuclease was supplied zygotically. The difference between the mutants and heterozygote controls was significant at P = 0.001 and P = 0.002 for the cases of zygotic and maternal-zygotic endonuclease, respectively. Some template disruptions also were seen in DmBlm mutants with only maternal endonuclease, but the difference was not significant relative to the heterozygous controls (P = 0.179). No cases of template disruption were seen in any of the controls, confirming that production of these events requires not only a mutant genotype at DmBlm, but a source of I-SceI endonuclease (Fig. 4A).

We characterized the lesions in nine of the template disruptions as shown in Fig. 4B. They included a variety of changes, most with deletions and some with more complex changes at the breakpoints. Two cases, TD2 and TD4, carried both molecular markers derived from the P{CaSpeR} chromosome (LC and RC), thus confirming that they were not the result of crossing over in flanking intervals. In two others, TD1 and TD6, there were flanking markers LA and/or RA, possibly indicating branch migration of one or both Holliday junctions.

Rad51 Mutant Rescues Recombinational Phenotype of DmBlm.

We showed previously (Fig. 3C Left) that DmBlm mutants have elevated mitotic recombination even without induced DSBs. This observation provided a simple way to test for interactions between genes. We measured the recombination frequency in male germ lines between markers at opposite ends of chromosome 2. As in all previous experiments, we used a series of single-male crosses to generate a set of independent measurements.

Fig. 5 shows the resulting recombination frequencies as measured in various genetic backgrounds. The most interesting comparison is between the first two rows, where both groups are mutant for DmBlm, but the second genotype also carries a mutation at spnA, the Drosophila Rad51 gene. This result indicates that the elevated recombination caused by DmBlm is abolished in the absence of Rad51 function. The sample size for the double mutants is small, owing to poor viability or fertility of the DmBlm-rad51 combination, but the difference is still significant at P < 10−6. This result parallels the finding by McVey et al. (29), who found that rad51 mutations rescued DmBlm deletion formation.

Fig. 5.
Male recombination frequencies for mutation combinations. Comparison of the first two rows shows that loss of the rad51 function abolishes the recombination induced by lack of DmBlm. The other rows confirm the effect on recombination by DmBlm and Top3α. ...

The remaining rows in Fig. 5 confirm the expected lack of recombination in Rad51 single mutants (row 3), the elevated recombination in DmBlm mutants (rows 4, 5, and 7), and a suggestion of elevated recombination in Top3α hypomorphs (row 6), although the last is not statistically significant (P = 0.09).


DSB Repair Outcomes Measured by Rr3.

Our measurements of SSA, NHEJ, and HR-h include all known pathways of DSB repair except those involving homologous repair from sister chromatid (HR-s) (Fig. 1A). Unlike the three measured outcomes, HR-s restores the I-SceI cut site, leaving the Rr3 construct available for additional cycles of breakage and repair, possibly in future cell generations. Therefore, SSA, NHEJ, and HR-h represent the terminal outcomes in a potentially cyclical process, and their measured frequencies provide inferences on the relative usage of the alternative pathways for repairing DSBs.

The relative frequencies of these three outcomes are measured in the germ line of each test male by scoring his offspring as shown in Fig. 1A. The sum of the three measurements need not be 100% because SSA is computed from the sons and the other two measures from the daughters. Furthermore, none of the three includes the “unchanged” class, which consists of uncut copies of Rr3 plus those which were cut but repaired by HR-s (see Supporting Methods). Empirically, however, the total tends to be close to 100%, as seen in the present data as well as previous work (24, 25). This tendency implies that nearly all copies of Rr3 are cut at least once in the germ-line lineage and repaired via SSA, NHEJ, or HR-h. Additional consideration of the statistical properties of these measures is in Supporting Methods.

Maternal Effects.

The genotype of the mutant DmBlm test males used in these experiments was mus309D2/Df(3)T7, which is null for DmBlm (26). However, some maternal-effect expression of BLM is possible early in the development of these males, because their mothers carried a WT DmBlm allele. This situation cannot be avoided owing to the infertility of DmBlm-null females (26). Evidence for the biological importance of this maternal effect contribution of DmBlm function can be seen in Fig. 4A, where the effect of the DmBlm mutation is shown to be minimal when the I-SceI endonuclease was supplied only via maternal effect. That is, the mutants behaved similarly to WT when DSB formation was limited to early development where maternal-effect BLM was still available. Moreover, when endonuclease was supplied both zygotically and maternally, the effect of the DmBlm mutation was actually less than when endonuclease was zygotic only (Figs. 3C and and44A). Our interpretation of this seemingly contradictory result is that the early DSBs formed by maternal endonuclease were repaired in the presence of maternal DmBlm, thus resulting in a phenotype closer to WT. These observations imply that the maternal component is a significant part of BLM function in Drosophila.

Genetic Dissection of DmBlm Roles.

Our data discern two cleanly separable functions of BLM in Drosophila. One is in repair of induced DNA breaks via homologous recombination with suppression of crossing over (Figs. 1 and and333–6).–6). The second BLM role is an essential function that can be rescued by a WT copy of mus81 (Fig. 2A). Other studies have implicated BLM and mus81 in maintaining replication forks (1719, 33, 38, 39). Our data are consistent with this idea, because mus81 itself has no detectable role in repairing induced DSBs (Figs. 2B and 9), even though the onset of double-mutant lethality resembles that of unrepaired DNA damage (Fig. 2A).

The Role of BLM in Homologous Repair.

Our data indicate that the loss of BLM function in the germ line of flies has the following effects: (i) a 4-fold decrease in HR-h with an offsetting increase in SSA; (ii) a 4-fold increase in deletions flanking DSB repair; (iii) a 50-fold or greater increase in mitotic recombination, especially where a DSB has been created; and (iv) the formation during DSB repair of template disruptions which do not occur at measurable frequencies in DmBlm+ flies. Owing to the likelihood, as discussed above, of maternal effect BLM function in the tested DmBlm mutants, we must assume that the extent of these effects is actually underestimated.

The first two observations listed above are expected if BLM has a role in HR either via the SDSA pathway or an alternative mechanism involving a dHJ intermediate, as reviewed by Symington (1). However, the remaining two observations are more consistent with BLM affecting repair via dHJs than SDSA. In fact, with certain assumptions listed below, all four observations can be explained by the proposal (11) that BLM functions in the dissolution of dHJ structures during homologous DSB repair. Our observation (Figs. 1 B and C and and5)5) that a weak allele of Top3α acts in the same direction as DmBlm, is consistent with this view, because both BLM and Top3α are needed in vitro for dissolution (11, 14). In addition, we find that loss of rad51 function, which would prevent the formation of dHJ structures, also precludes recombination in DmBlm mutants (Fig. 5). According to this model, our observations of increased recombination (Figs. 1B, ,44C, and and5),5), flanking deletions (Fig. 1C), and template disruptions (Fig. 4) are explained by the occurrence of new DSBs formed when dHJ structures cannot undergo dissolution. These new breaks would be expected in one or both of the duplexes and would then have to be repaired by NHEJ or another method, resulting in deletions and/or recombinants.

The results in Figs. 1B and and44A show that loss of zygotic BLM abolishes most HR-h events. Therefore, if the sole function of BLM in repairing induced DSBs is the dissolution of dHJ structures, it implies that most HR-h events do involve a dHJ intermediate. Counter to this idea are previous arguments (40) that SDSA, which does not invoke dHJ intermediates, is the most common pathway for homologous repair in Drosophila premeiotic germ cells. However, some of the evidence for SDSA being the dominant pathway was based on the dearth of DSB repair-associated crossing over or alterations in the template sequence, both of which can also be explained by the formation and frequent dissolution of dHJ structures. Adams et al. (21) found that DmBlm mutants showed a marked decrease in a class of truncated repair events in which a direct duplication present on the template is collapsed in the repaired chromosome. They interpret these events in terms of SDSA repair. This interpretation is complicated, however, by the requirement in their experimental system for filling in of a 14-kb gap. Successful repair of gaps >11 kb occurs only at a much reduced frequency (41) and might involve atypical events.

The template disruptions we observed (Fig. 4) are much more easily explained by a breakdown in dHJ resolution rather than a defect in SDSA, because the latter does not normally involve strand exchange in the template duplex (1, 42). These template disruption events might be analogous to an observation by Seki et al. (43) who found cases in mammalian cells deficient for Top3α in which both sister chromatids underwent breakage in the same vicinity. They interpreted these events in terms of failure of the dissolution step to separate chromatids.

We conclude that our data are consistent with dissolution of dHJ structures being the primary, or even sole, role of BLM in the homologous repair of induced DSBs, but other roles are not precluded. In addition, the synthetic lethality with mus81 implies at least one other role for BLM, perhaps in the recovery from replication fork collapse.


Drosophila rearing conditions and the use of Rr3 to measure parameters of DSB repair were as per Preston et al. (24) with modifications described in Figs. 1 and and2.2. Additional details and examples are available in Supporting Methods. Tests of methyl methanesulfonate sensitivity used a 0.1% mutagen solution, as described in Supporting Methods and ref. 28. Test crosses with genetic markers listed in Supporting Methods were used to compute template disruption and long-gap HR-h frequencies.

Standard errors were computed for all measurements by using single-male replicate data as described in Results. All hypothesis tests were performed by using permutation tests in which each single-male replicate was taken as an independent observation. Further details and examples are in Supporting Methods.

Supplementary Material

Supporting Information:


Christine Preston and Carlos Flores provided much appreciated help and advice throughout this project, and Ann DeLaForest and Christie Miller provided technical help. This work was supported by National Institutes of Health Grant GM30948.


dHJdouble Holliday junction
DSBdouble-strand break
HRhomologous repair
HR-hhomologous repair from homolog
HR-shomologous repair from sister chromatid
NHEJnonhomologous end-joining
SDSAsynthesis-dependent strand annealing
SSAsingle-strand annealing.


The authors declare no conflict of interest.


1. Symington LS. Microbiol Mol Biol Rev. 2002;66:630–670. [PMC free article] [PubMed]
2. Cahill D, Connor B, Carney JP. Front Biosci. 2006;11:1958–1976. [PubMed]
3. Valerie K, Povirk LF. Oncogene. 2003;22:5792–5812. [PubMed]
4. Fishman-Lobell J, Rudin N, Haber JE. Mol Cell Biol. 1992;12:1292–1303. [PMC free article] [PubMed]
5. Moore JK, Haber JE. Mol Cell Biol. 1996;16:2164–2173. [PMC free article] [PubMed]
6. Bachrati CZ, Hickson ID. Biochem J. 2003;374:577–606. [PMC free article] [PubMed]
7. Chaganti RS, Schonberg S, German J. Proc Natl Acad Sci USA. 1974;71:4508–4512. [PMC free article] [PubMed]
8. Ellis NA, Groden J, Ye TZ, Straughen J, Lennon DJ, Ciocci S, Proytcheva M, German J. Cell. 1995;83:655–666. [PubMed]
9. Cheok CF, Wu L, Garcia PL, Janscak P, Hickson ID. Nucleic Acids Res. 2005;33:3932–3941. [PMC free article] [PubMed]
10. Karow JK, Chakraverty RK, Hickson ID. J Biol Chem. 1997;272:30611–30614. [PubMed]
11. Wu L, Hickson ID. Nature. 2003;426:870–874. [PubMed]
12. Wu L, Bachrati CZ, Ou J, Xu C, Yin J, Chang M, Wang W, Li L, Brown GW, Hickson ID. Proc Natl Acad Sci USA. 2006;103:4068–4073. [PMC free article] [PubMed]
13. Wu L, Chan KL, Ralf C, Bernstein DA, Garcia PL, Bohr VA, Vindigni A, Janscak P, Keck JL, Hickson ID. EMBO J. 2005;24:2679–2687. [PMC free article] [PubMed]
14. Plank JL, Wu J, Hsieh TS. Proc Natl Acad Sci USA. 2006;103:11118–11123. [PMC free article] [PubMed]
15. Leng M, Chan DW, Luo H, Zhu C, Qin J, Wang Y. Proc Natl Acad Sci USA. 2006;103:11485–11490. [PMC free article] [PubMed]
16. Opresko PL, Mason PA, Podell ER, Lei M, Hickson ID, Cech TR, Bohr VA. J Biol Chem. 2005;280:32069–32080. [PubMed]
17. Franchitto A, Pichierri P. J Cell Biol. 2002;157:19–30. [PMC free article] [PubMed]
18. Ralf C, Hickson ID, Wu L. J Biol Chem. 2006;281:22839–22846. [PubMed]
19. Sengupta S, Robles AI, Linke SP, Sinogeeva NI, Zhang R, Pedeux R, Ward IM, Celeste A, Nussenzweig A, Chen J, et al. J Cell Biol. 2004;166:801–813. [PMC free article] [PubMed]
20. Min B, Weinert BT, Rio DC. Proc Natl Acad Sci USA. 2004;101:8906–8911. [PMC free article] [PubMed]
21. Adams MD, McVey M, Sekelsky JJ. Science. 2003;299:265–267. [PubMed]
22. Rockmill B, Fung JC, Branda SS, Roeder GS. Curr Biol. 2003;13:1954–1962. [PubMed]
23. Bachrati CZ, Borts RH, Hickson ID. Nucleic Acids Res. 2006;34:2269–2279. [PMC free article] [PubMed]
24. Preston CR, Flores CC, Engels WR. Genetics. 2006;172:1055–1068. [PMC free article] [PubMed]
25. Preston CR, Flores CC, Engels WR. Curr Biol. 2006 Oct 24; doi: 10.1016/j.cub.2006.08.058. [PubMed] [Cross Ref]
26. Kusano K, Johnson-Schlitz DM, Engels WR. Science. 2001;291:2600–2602. [PubMed]
27. Kusano K, Berres ME, Engels WR. Genetics. 1999;151:1027–1039. [PMC free article] [PubMed]
28. Boyd JB, Golino MD, Shaw KE, Osgood CJ, Green MM. Genetics. 1981;97:607–623. [PMC free article] [PubMed]
29. McVey M, LaRocque JR, Adams MD, Sekelsky JJ. Proc Natl Acad Sci USA. 2004;101:15694–15699. [PMC free article] [PubMed]
30. Plank JL, Chu SH, Pohlhaus JR, Wilson-Sali T, Hsieh TS. J Biol Chem. 2005;280:3564–3573. [PubMed]
31. Grumbling G, Strelets V. Nucleic Acids Res. 2006;34:D484–D488. [PMC free article] [PubMed]
32. Boddy MN, Lopez-Girona A, Shanahan P, Interthal H, Heyer WD, Russell P. Mol Cell Biol. 2000;20:8758–8766. [PMC free article] [PubMed]
33. Boddy MN, Gaillard PH, McDonald WH, Shanahan P, Yates JR III, Russell P. Cell. 2001;107:537–548. [PubMed]
34. Haber JE, Heyer WD. Cell. 2001;107:551–554. [PubMed]
35. Mullen JR, Kaliraman V, Ibrahim SS, Brill SJ. Genetics. 2001;157:103–118. [PMC free article] [PubMed]
36. Ashburner M. Plainview, NY: Cold Spring Harbor Lab Press; 1989. Drosophila, A Laboratory Handbook.
37. Ii M, Brill SJ. Curr Genet. 2005;48:213–225. [PMC free article] [PubMed]
38. Gao H, Chen XB, McGowan CH. Mol Biol Cell. 2003;14:4826–4834. [PMC free article] [PubMed]
39. Abraham J, Lemmers B, Hande MP, Moynahan ME, Chahwan C, Ciccia A, Essers J, Hanada K, Chahwan R, Khaw AK, et al. EMBO J. 2003;22:6137–6147. [PMC free article] [PubMed]
40. Nassif N, Penney J, Pal S, Engels WR, Gloor GB. Mol Cell Biol. 1994;14:1613–1625. [PMC free article] [PubMed]
41. Johnson-Schlitz DM, Engels W. Genetics. 2006;173:2033–2038. [PMC free article] [PubMed]
42. Formosa T, Alberts BM. Cell. 1986;47:793–806. [PubMed]
43. Seki M, Nakagawa T, Seki T, Kato G, Tada S, Takahashi Y, Yoshimura A, Kobayashi T, Aoki A, Otsuki M, et al. Mol Cell Biol. 2006;26:6299–6307. [PMC free article] [PubMed]
44. Preston CR, Sved JA, Engels WR. Genetics. 1996;144:1623–1638. [PMC free article] [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem chemical compound records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records. Multiple substance records may contribute to the PubChem compound record.
  • Gene
    Gene records that cite the current articles. Citations in Gene are added manually by NCBI or imported from outside public resources.
  • GEO Profiles
    GEO Profiles
    Gene Expression Omnibus (GEO) Profiles of molecular abundance data. The current articles are references on the Gene record associated with the GEO profile.
  • HomoloGene
    HomoloGene clusters of homologous genes and sequences that cite the current articles. These are references on the Gene and sequence records in the HomoloGene entry.
  • MedGen
    Related information in MedGen
  • Pathways + GO
    Pathways + GO
    Pathways and biological systems (BioSystems) that cite the current articles. Citations are from the BioSystems source databases (KEGG and BioCyc).
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.
  • Taxonomy
    Taxonomy records associated with the current articles through taxonomic information on related molecular database records (Nucleotide, Protein, Gene, SNP, Structure).
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...